Potsdam, Germany 21.07.06 Identification and removal of glitches in GEO 600 Joshua Smith ILIAS WG1 21.07.06.

Potsdam, Germany 21.07.06

Identification and removal Identification and removal of glitches in GEOof glitches in GEO 600600

Joshua Smith

ILIAS WG1

21.07.06

Joshua Smith, Potsdam, Germany

21.07.06

OverviewOverview

Tools:o GEO Summary pageso mHACRo Coincident analysiso Noise projections

Glitches:o Owlo Dusto Frequency glitcheso MID slow FB glitcheso Magnetic field/mains

Since start of S5 emphasis has shifted to data quality

- works by: making lots of automatic characterization, mining this for information, finding source

Joshua Smith, Potsdam, Germany

21.07.06

mHACRmHACR• Time-frequency map of excess power (based on HACR)

• Used to search for burst-like events

• Matlab routine run on cluster (Martin & Ajith)

• whitens data, removes lines

• writes events to database

• queried by mysql/matlab

HACR: what’s that??

-Heating, Air Conditioning & Refrigeration

-Helicopter Active Control Rotor

-Hierarchical Algorithm for Curves and Ridges

-High Accuracy Cryogenic Radiometer

-Hispanic Association on Corporate Responsibility

Joshua Smith, Potsdam, Germany

21.07.06

Coincident analysisCoincident analysis

Joshua Smith, Potsdam, Germany

21.07.06

OwlOwl

Joshua Smith, Potsdam, Germany

21.07.06

Owl gone???Owl gone???

Not seen after work on the signal-recycling longitudinal error signal moduleo Broken switch from one error signal to anothero This can’t be the cause of the owl, since owl

existed before this module

Gone for good?

Joshua Smith, Potsdam, Germany

21.07.06

DustDust

Joshua Smith, Potsdam, Germany

21.07.06

Power glitchesPower glitches

Detector output

Input port power

Joshua Smith, Potsdam, Germany

21.07.06

Back-coupling?Back-coupling?

Strain coupling to DC power

Dark port DC Power

Strain back-coupling doesn’t explain instrumental channel coincidence

Projection from detector output to laser power

Joshua Smith, Potsdam, Germany

21.07.06

Power glitches vs. dustPower glitches vs. dust

Dark port DC power glitch count

Dark port DC power glitch count

Particle count

Dark port DC power glitch count

Input port DC power glitch count

0 Time [m] 180

Joshua Smith, Potsdam, Germany

21.07.06

Particle waves filteredParticle waves filtered14020

455

30.2

(Par

ticl

es/c

u.ft

)/10

00

A

a

Joshua Smith, Potsdam, Germany

21.07.06

Dust made loudest glitchesDust made loudest glitches

Before filter After filter

Time [h] Time [h]Det. output BLRMS @~ 170Hz

Joshua Smith, Potsdam, Germany

21.07.06

Frequency glitchesFrequency glitches

Joshua Smith, Potsdam, Germany

21.07.06

CharacteristicsCharacteristics0s 60s •Shows up most strongly in

PR FB

•Couples via frequency noise (PR EP)

Joshua Smith, Potsdam, Germany

21.07.06

Removing glitchesRemoving glitches

• Traced main glitches to a saturation somewhere in modecleaner2 loop.

• Drastically improved with 2x reduction of RF modulation depth (compensated by digital gain)

Joshua Smith, Potsdam, Germany

21.07.06

PR setupPR setup

Joshua Smith, Potsdam, Germany

21.07.06

Before and afterBefore and after

PR loop feedback

H

Joshua Smith, Potsdam, Germany

21.07.06

Glitches back, a mysteryGlitches back, a mystery

HACR plots from last week

Joshua Smith, Potsdam, Germany

21.07.06

Magnetic fields / MainsMagnetic fields / Mains

Joshua Smith, Potsdam, Germany

21.07.06

Magnetic field glitchesMagnetic field glitches

B glitches H glitches

Linear coupling, coherence

Joshua Smith, Potsdam, Germany

21.07.06

Magnetic field “projection”Magnetic field “projection” B-field nearly limiting noise around 100Hz?

o Coupling not yet understood o Checks: magnetometer measures B-field, IFO back-coupling ruled out

Joshua Smith, Potsdam, Germany

21.07.06

Hourly mains glitchesHourly mains glitches

Hourly series start:o 24 secs after GPS houro 10 secs after UTC hour

24s

Joshua Smith, Potsdam, Germany

21.07.06

Mag./mains coinc. with HMag./mains coinc. with H

Fr e

quen

c y [

Hz]

Magnetometer central building ~ 50coinc./8hrs

Time-freq glitch maps

Time-shifted coincidence

Mains monitor central building ~ 30coinc./8hrs

Time [h] Time [h]0 08 8

50

0

50

0

Time shift [s] Time shift [s]1 1-1 -1

Fre

quen

c y [

Hz]

Joshua Smith, Potsdam, Germany

21.07.06

MID slow FB glichesMID slow FB gliches

Joshua Smith, Potsdam, Germany

21.07.06

CharacteristicsCharacteristics

Looks like a switch!

Joshua Smith, Potsdam, Germany

21.07.06

Coincident with mains glitchesCoincident with mains glitches

Joshua Smith, Potsdam, Germany

21.07.06

HACR History PlotsHACR History PlotsMay June

July

Joshua Smith, Potsdam, Germany

21.07.06

Loose endsLoose ends

• Is the owl gone?

• Is the dust concentration low enough?

• Why did frequency glitches come back?• Are these related to power-up glitches?

• How to solve mains glitches?

• Are slow FB glitches caused by a switch?

• Q for all: Glitch vs. Instability??

Joshua Smith, Potsdam, Germany

21.07.06

The endThe end

'falco tinnunculus',

better know as 'Turmfalke'

Joshua Smith, Potsdam, Germany

21.07.06

Statistical dust veto?Statistical dust veto?

MID VIS veto TF map and timeshifted results